re-use initial population for multiple trials?

Hi, suppose I would like to run a GA to solve the TSP
problem. I want to execute my program 100 times to find
the average best result across the 100 executions
(where each execution has lots of generations, say 500).
Should I start with a different initial population for each
of the runs? I figure that it does not matter since the
stochastic recombinations and mutations will make
the initial population moot, but I am not sure.

.

Relevant Pages

  • Re: re-use initial population for multiple trials?
    ... > want to execute my program 100 times to find the average best result ... > initial population moot, ... Unless you are comparing your algorithm against another algorithm i ... There are three types of posters to alt.binaries.* ...
    (comp.ai.genetic)
  • Run code on new record
    ... not execute, making me think that my plan is no good. ... Prev by Date: ...
    (microsoft.public.access.modulesdaovba)
  • Run after deployment
    ... I'm looking for a quick way to execute my app immidiately after its ... I'm deploying my app using VS2003 setup project. ... Prev by Date: ... Next by Date: Unable to open the Web 'http://localhost', FrontPage Server extensions not installed ...
    (microsoft.public.dotnet.general)
  • Re: VS2005 seems SLOOOOOW
    ... > my program from within the IDE? ... Prev by Date: ...
    (microsoft.public.vsnet.general)
  • Re: nieuws gebruikers
    ... Enquire, plan and execute. ... Stourport, England ... Prev by Date: ...
    (microsoft.public.windows.inetexplorer.ie6_outlookexpress)